267: Cute and Squishy

Embedded.fm · 9 Nov 2018 · 1h 04m · Episode 267

What this episode covers

Lindsey Kuper, a new professor of computer science and engineering at UC Santa Cruz and co-founder of Bang Bang Con, discusses her research on programming languages and distributed systems, the challenges of reading technical papers, and the logistics of organizing the Bang Bang Con West conference. She explains the trade-offs between expressivity, efficiency, and generality in language design, citing Julia and GPUs as examples, and details the specific requirements for talk proposals at the upcoming Santa Cruz event.

Mentioned

Companies
Apnexus
Also
Bang Bang Con, UC Santa Cruz, Julia, Recurse Center, Leslie Lamport, ICFP, Starkon

Other shows on the same ground, closest first.

  • Oxide and Friends · 1h 26m

    Dijkstra's Tweetstorm

    Bryan Cantrill and Adam Leventhal of Oxide Computer Company discuss Edsger Dijkstra’s 1975 tweetstorm, using it to explore how programming languages shape cognitive models and engineering decisions. The conversation covers the pedagogical value of early languages like BASIC and Perl, the specific challenges of learning Rust, and the hardware bring-up struggles at Oxide involving iCE40 FPGAs and I2C level translators. They emphasize that while early languages like BASIC foster enthusiasm, deeper understanding requires exposure to lower-level concepts like C and memory management, a lesson reinforced by Oxide's own hardware debugging experiences.

  • Oxide and Friends · 29m

    golang asserts and the PLATO terminal

    Bryan Cantrill, Adam Jacob, and other engineers discuss the philosophy of programming languages, specifically contrasting the strict, opinionated design of Go and Elm with the freedom of Rust and Smalltalk. The conversation shifts to the history of the PLATO terminal system, with Bob Mader and Tom Lyon sharing personal experiences using the system in the 1970s, including its plasma displays and TUTORDart programming. They highlight Brian Dear's book 'The Friendly Orange Glow' as a key resource for understanding PLATO's advanced features and social dynamics, comparing it to the Xerox Alto and discussing the legacy of Dr. David Gräper's Grapenotes.

  • Oxide and Friends · 1h 33m

    Pragmatic LLM usage with Nicholas Carlini

    Security researcher Nicholas Carlini joins Bryan Cantrill and Adam Leventhal to discuss his blog post on the pragmatic use of Large Language Models (LLMs) as tools for reducing friction in software development. Carlini shares specific examples of using GPT-4 and Claude 3.5 Sonnet to simplify code, decompile Python bytecode, and generate plots, while emphasizing the importance of verifying outputs and understanding the models' limitations. The conversation also explores the impact of LLMs on education, hiring practices, and the potential for these tools to democratize programming by lowering the barrier to entry for beginners.

  • Oxide and Friends · 1h 17m

    No Silver Bullets

    Bryan Cantrill and Steve Klabnik discuss Fred Brooks' essay 'No Silver Bullet,' arguing that while technologies like Rust, Rails, and open source collaboration eliminate accidental complexity, they cannot remove the essential complexity of software problems. They analyze Brooks' distinction between essential and accidental complexity, noting that tools like Rust force developers to confront the latter, while platforms like S3 and the Internet enable massive productivity gains through better collaboration rather than magic bullets. The conversation also covers Brooks' skepticism regarding AI and graphical programming, concluding that no single technology offers a 10x or 100x improvement because the core difficulty of defining what to build remains unchanged.

  • Oxide and Friends · 1h 29m

    AI in Computer Science Education

    Brown University computer science professors Shriram Krishnamurthi and Kathi Fisler discuss their experimental 'AgenTic Studio' course, which integrates agentic programming tools like Claude into the curriculum. They detail their pedagogical approach of treating LLMs as 'flaky compilers' and using assignments like generating Tetris variants to teach students about testing, types, and the limits of AI-generated code. The conversation covers student reactions, the challenge of grading, and the future of CS education as institutions adapt to this new reality.

The summary, chapters and topics are generated from the episode’s audio by this directory, not written by Embedded.fm. The episode itself belongs to its publisher.